gpt-oss-safeguard-20b
gpt-oss-safeguard-20b is a large language model made by OpenAI, added to the catalog on 29 Oct 2025. It has a context window of 131,072 tokens and costs $0.075 per 1M input tokens and $0.30 per 1M output tokens. It accepts text and produces text. Typical uses: enhance content safety classification, improve LLM filtering processes and streamline trust and safety labeling.
Price checked · source: OpenRouter
What gpt-oss-safeguard-20b is
The OpenAI gpt-oss-safeguard-20b, released in October 2025, is a text-to-text model designed for safety reasoning tasks. Built on the gpt-oss-20b framework, this model features a 21 billion parameter Mixture-of-Experts architecture, providing efficient performance for tasks such as content classification, LLM filtering, and trust & safety labeling. With a substantial context window of 131,072 tokens, it supports various text input and output modalities. This model is particularly suited for applications requiring nuanced safety evaluations, offering lower latency in processing complex safety-related queries.
Key features
- 21B-parameter Mixture-of-Experts architecture
- Designed for safety reasoning tasks
- Supports content classification and LLM filtering
- Large context window of 131,072 tokens
- Text-to-text input and output modalities
Typical uses
- Enhance content safety classification
- Improve LLM filtering processes
- Streamline trust and safety labeling
- Optimize safety-related query processing
gpt-oss-safeguard-20b pricing
US dollars, list price through OpenRouter.
| Item | Standard |
|---|---|
| Input, per 1M tokens | $0.075 |
| Output, per 1M tokens | $0.30 |
| Cached input read, per 1M tokens | $0.0375 |
No price change since it was added on 29 Oct 2025.
| What a task costs | Cost |
|---|---|
| Answer a short question500 in / 300 out | $0.0001 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.0021 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.0029 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.012 |
gpt-oss-safeguard-20b specifications
| Model id | openai/gpt-oss-safeguard-20b |
|---|---|
| Version | gpt-oss-safeguard-20b |
| Context window | 131,072 tokensAbout 98,304 English words |
| Max output | 65,536 tokens |
| Knowledge cutoff | Not published |
| Added to catalog | 29 Oct 2025 |
| Input types | text |
|---|---|
| Output types | text |
| Tool calling | Yes |
| Structured outputs | Yes |
| Reasoning effort control | No |
| Content moderation | Not applied by provider |
Sample outputs
The same prompts are given to every model, and the unedited output is shown with the date it was produced. Upend does not publish leaderboard scores it did not run.
Write a short essay
RUN20 Nov 2026Write 150 words on how AI might positively upend work, leisure and creativity
Artificial intelligence is poised to reshape work, leisure, and creativity in ways that amplify human potential rather than merely replace it. In the workplace, AI can automate routine tasks, freeing employees to focus on strategic, empathetic, and innovative work. Predictive analytics and intelligent assistants enable smarter decision‑making, while collaborative robots (cobots) augment physical labor, improving safety and productivity. For leisure, AI curates personalized experiences—music, art, and travel—by learning individual preferences and generating novel content in real time. It also powers immersive virtual environments that blend gaming, education, and social interaction, making downtime more enriching. Creatively, AI acts as a co‑creator, offering instant drafts, suggesting novel combinations, and democratizing access to sophisticated tools. By lowering technical barriers, it empowers amateurs to produce professional‑grade art, music, and literature. Together, these shifts foster a culture where human ingenuity is amplified, work is more meaningful, and leisure becomes a fertile ground for exploration, everywhere, always.